[EM] FairVote comment on Burlington dumping IRV

[Abd ul-Rahman Lomax]

http://www.fairvote.org/lessons-from-burlington#.UdOvX2LE0XY (March 4, 2010)

>Let me cut to the chase. Despite winning in five 
>of the city’s seven wards, the use of instant 
>runoff voting (IRV) for mayor was repealed this 
>week by a margin of less than 4% in Vermont’s 
>largest city of Burlington. It’s a 
>disappointment, particularly with a growing 
>appreciation in Vermont for IRV. Those strongly 
>opposing repeal of IRV included the state’s 
>leading civic groups – VPIRG and state arms of 
>the League of Women Voters and Common Cause – 
>and a host of political leaders, including Sen. 
>Bernie Sanders, Gov. Howard Dean and nearly 
>every state legislator from the city.

He's got to lead with all the reasons why 
Burlington voters were stupid. FairVote convinced 
the LWV and CC to adopt a pro-IRV stand many 
years ago, and it turns out that these 
organizations are a bit like FairVote. They don't 
reconsider, they just keep on keeping on, once a cause has been decided.

>IRV in Burlington only has been used in two 
>mayoral races, as it was not adopted for city 
>council races. Its defeat stems from a simple 
>fact: the only candidate ever to win with IRV in 
>Burlington is current mayor Bob Kiss, who won 
>two elections for mayor in 2006 and 2009 in 
>hotly contested races where no candidate won 40% 
>of first choices. In a city with three major 
>parties, all with roughly comparable support, 
>victories for only one party's nominee meant 
>that a majority of voters had yet to see their 
>first choice win in an IRV race. Kiss was the 
>majority choice over his top opponents in 2006 
>and 2009, to be sure, but with new controversies 
>in the his administration in the past year, it 
>was clear that in a referendum on the mayor this year, he would lose.

In other words, "not our fault, not IRV's fault. The problem is Kiss."

However, "Kiss was the majority choice over his 
top opponents in 2006 and 2009"?

In 2009, Kiss had 48% of the vote. IRV creates a 
false majority by eliminating candidates, it's 
called a "last round majority." It excludes all 
the voters who voted against the top two. Richie 
is so accustomed to saying that IRV finds 
majorities that he may even believe it.

>Opponents of IRV were well aware of this fact, 
>and did everything they could to attach IRV to 
>Kiss. At a televised debate, they carried signs 
>saying “Where’s Bob,” suggesting Kiss should be 
>the one defending IRV. They called out at the 
>polls “if you don’t like Bob Kiss, vote to 
>repeal IRV.” They focused on the 2009 election 
>results, suggesting that IRV had cheated voters 
>into a Kiss victory so that backers would have 
>to explain how in fact Kiss had earned his 
>majority win. After Tuesday’s vote, one city 
>councilor called on Kiss to resign – showing the 
>direct link in many voters’ minds between IRV and the mayor.

OMG, opponents made *political arguments*! 
FairVote has sold itself as politically savvy. 
Who would those "calling out" be talking to? How 
about the majority of voters who opposed Kiss in 
the election? It would especially be the 
Republicans, with the largest single party 
(because of the Democrats and Progressives being separate).

Yeah, I know there is something particularly 
scuzzy about *appealing to the majority.* But ... 
wasn't IRV supposed to find majority results? 
Wasn't that the very point? The very fact that 
Richie thinks that this argument was telling shows that IRV failed.

Really unfair, sure. They focused on the 2009 
election results! How illogical to judge a voting 
system by its results! Don't they know that IRV 
has "momentum"? Never mind that voting systems 
experts anticipated failures like that. Totally 
predictable -- i.e., quite likely -- in 
situations as in Burlington, with three major 
parties. Looking at first preference votes in 
2009: Republican: 2951, Progressive 2585, Democrat 2063.

>IRV opponents were led by Kurt Wright, who lost 
>the 2009 race in a cliffhanger. Wright had led 
>after the count of first choices and continued 
>to lead in the count until the field was reduced 
>to two. In the final instant runoff, a majority 
>of voters ranked Kiss ahead of Wright, giving 
>Kiss re-election. Within weeks of his defeat, 
>Wright's supporters were in the streets 
>collecting petitions for repeal – joined by some 
>backers of other losing candidates. Their drive 
>seemed to falter after initial enthusiasm, but 
>then a public scandal enveloped the mayor, and 
>petition gatherers rushed to finish getting 
>their repeal on the ballot. In the repeal, the 
>two wards where Wright ran most strongly voted 
>against IRV by a margin of two-to-one after 
>supporting it when first passed in 2005. The 
>rest of the city voted 60% to keep IRV.

Richie is, so far, not acknowledging *any 
problems* with IRV. It's all politics. Yet *the 
majority of voters* in 2009 voted *against* Kiss. 
In that "final instant runoff," there were many 
ballots no longer counted because every candidate 
other than Kiss and Wright, including the obvious 
compromise candidate, the Democrat, had been 
eliminated. In 2012, the Democrat won, the 
Progressives not having fielded a candidate.

Notice how every argument he can come up with the 
make it seem that something was wrong with the 
repeal. But it would not at all be surprising for 
Wright supporters to be upset about IRV. The key 
would be if they could enlist the support of Democrats. Obviously, they did.

It's very true that with the system they went 
back to, top-two runoff with a 40% margin 
required, in 2009, Kiss might have won, though 
that's not entirely clear. (people might vote 
differently, surprises happen in runoffs.)

>Reformers can’t control who wins and loses 
>elections, but the lesson from Burlington is 
>they need to be aware that many voters measure 
>the value of a reform by who wins under the new 
>rules. In the case of Burlington, exit polls 
>after the first IRV election in 2006 found 
>overwhelming support for IRV, with voters four 
>times more likely to support it than oppose it 
>and only a handful saying they found it 
>confusing. But that was before they knew who 
>won. As soon as you have winners and losers, as 
>of course you always will, some voters will 
>rethink their assessment. If the same candidate 
>wins twice in a city where that candidate 
>commands perhaps a third of the vote, you have to be ready.

Bingo. However, it's not that simple. "commands 
perhaps a third of the vote" refers to first 
preference, and many voters don't think like 
political activists. They don't like to be 
"commanded." The problem is that IRV *pretends* 
to be a majority system, when it is very often a 
plurality system, it elects when a majority of 
voters have voted *against* the candidate. If IRV 
attempted to find a majority, but if there were a 
real runoff if it fails to do so, matters might 
be different. But we wouldn't use IRV for this, 
we'd use Bucklin. It's much better at finding 
real majorities, because it does not eliminate 
candidates and thus disregard the ballots of 
those who voted only for them. Those remain part of the basis for "majority."

IRV effectively lies about "majority." And people 
very much did not like that, when they discovered 
it. Did they only vote on Kiss and not on IRV? 
Well, does it matter? By electing a candidate who 
was not the true majority choice (that would be 
Montroll), IRV set up conditions for repeal.

>This helps explain that keeping reform can 
>sometimes be harder than winning it – at least 
>until it’s understood as a change that doesn’t 
>favor one side. The irony is that because IRV 
>has been such a potent electoral vehicle -- 
>winning by landslide margins in ballot measures 
>in a range of cities such as Memphis (TN), 
>Oakland (CA) and Minneapolis (MN). that it can 
>be won before there is much grassroots effort to 
>introduce it to voters. There’s a gut appeal to 
>winning majorities in one round of voting and to 
>being able to rank candidates instead of just 
>“X”-voting for them. That’s good for IRV, but 
>dangerous when those first results come in, and 
>backers of losing candidates finger the new rule 
>as the reason their favorite candidate lost.

Let's not mention the other places which tried it 
and dropped it. "Gut appeal to winning majorities 
in one round of voting" is an appeal to an error. 
IRV cannot do that, no voting system can 
guarantee that *except by disregarding votes or 
excluding all but two candidates and disallowing 
write-ins.* Richie just keeps repeating his 
mantra. Who is reading him any more? Perhaps the 
President of the Arizona League of Women Voters, 
who happens to be the President of Arizona FairVote.

>Local backers [www.fiftypercentmatters.com] did 
>a terrific job in responding to the attack on 
>IRV, although formed their campaign too late to 
>fully dispel some of the rampant misconceptions 
>being spread by opponents. I hope that education 
>is done because Burlington remains a city with a 
>strong coalition that backs IRV and a problem to 
>fix. The city has three major parties and a 
>history of independent candidates, and the 
>system just voted in – one allowing a candidate 
>to be elected with just 40% of the vote – brings 
>back the problems of “spoilers” and minority rule.

That's weird. IRV behaves like top-two runoff, 
that is why it was named "instant runoff," except 
that it doesn't allow voters that second look. 
Richie apparently has a pile of 'standard 
arguments' that he pops into every piece, a 
collection of slogans and sound-bites. "Minority 
rule" ... but IRV allowed minority rule. Top-two 
runoff, if write-ins are allowed in the runoff, 
can be superior to IRV. There are much better 
ways, but FairVote has attempted to discredit 
them, with extensive propaganda. Many years ago, 
they convinced the LWV to back IRV, with 
propaganda about approval voting, which is really 
a *very simple reform*, not the ideal but simply 
fixing a problem, allowing voters to vote for 
more than one *if they choose*. There are 
definitely better methods, but this one requires 
no change to ballots except "vote for as many as 
you choose, the winner will be the candidate with 
the most votes." And that's already a standard 
voting method where there are multiple 
conflicting ballot questions, which we could 
think of a many-candidate election as being.

>More broadly, backers of IRV in other cities and 
>states adopting IRV must work to keep reminding 
>people how the system works and to be clear in 
>explaining the results after they happen, 
>starting with the media (which did an inadequate 
>job in Burlington after the 2009 election, 
>leading some to think that some voters had two 
>votes and others just one). New cities facing 
>that opportunity this year included Berkeley, 
>Oakland and San Leandro in California.

Basically, it's not enough to lie to them at the 
beginning. You have to *keep lying to them.* What 
is this "two votes" thing? He doesn't cite it.

>Going even more broadly, that same education 
>challenge exists for all reformers – never take 
>your wins for granted, and keep working to make 
>the case for your change even after you win it. 
>You will reach a point where your proposal is 
>the accepted status quo – for IRV backers, that 
>will happen with the idea of voting becomes 
>ranking. At that point your reform probably is 
>safe, as it seems to be in other countries using 
>IRV for decades such as Australia and Ireland.

The problem is that Richie and FairVote take the 
*value of IRV* for granted. It never occurs to 
them that there are problems, they reject all 
other proposed reforms, FairVote activists argue 
that everyone else should simply accept IRV 
because there is no other practical reform that 
might be accepted. Yet IRV is a suicidal reform, 
in the U.S. system. In Australia, voters are 
*required to rank all the candidates*, in most 
places. The history of IRV in Australia is 
interesting. It was introduced by a major party 
to disempower an expected spoiler, as I recall. 
IRV does fix that spoiler effect, when the minor 
party is small. But if it grows, it can then fail 
as IRV did in Burlington. Richie has not 
acknowledged, *at all*, that there was any problem with the IRV results.

Previously, voting systems experts used to 
criticize IRV, describing scenarios like that, 
and Richie pooh-poohed it as impossible. When it 
actually happens, he looks everywere except at 
*what actually offended the voters.*

Maybe Burlington was a bad place to suggest IRV!

Bucklin would work very well there. But FairVote 
is a one-trick pony. Yes, they supposedly want 
proportional representation, but they have put 
their major campaign effort into IRV, a very poor 
single-winner method, it's been known as such since the 19th century.

>What American reformers have going for us is 
>that the constraints of our two-choice system 
>are bitterly resented by a growing number of 
>Americans. IRV represents a means to accommodate 
>more options and encourage more inclusive modes 
>of campaigning in a range of settings, such as 
>nonpartisan elections and primaries. The case 
>for IRV remains as strong as ever, and 
>appreciation for its value keeps expanding. 
>Losses hurt and lessons from them must not be 
>forgotten, but our nation’s shift to “rank the vote” continues.

FairVote has papered over the history of Bucklin, 
with misleading propaganda. It was called, in the 
1920s, "American Preferential Voting." It was 
implemented in something like ninety cities and 
towns. It worked. It was dumped for -- almost 
everywhere -- purely political reasons, having 
nothing to do with *voters* being dissatisfied 
with results. It would work perfectly in 
Burlington. It is "instant runoff approval 
voting." It is simple to understand and canvass.

(Bucklin was last used in party primary elections 
in the South, and, FairVote reports that "only 
11%" of voters, in some elections, added 
additional preferences. That would happen with 
IRV, the same. FairVote claims that voters don't 
add additional votes because these could "hurt 
their favorite," but that is not so likely to 
happen in a party primary election, where voters 
do have a motive to choose the most 
widely-supported candidate. They do, after all, 
want to win the ultimate election for their 
party. The fact is, though, that many voters 
don't know more than one candidate, it's a very 
old problem, very well known. Only if you have a 
runoff in elections like that is it reasonably 
likely that a majority will be found. Bucklin 
was, by the way, replaced with top-two runoff. It 
would have been better if Bucklin had been 
modified to hold a runoff if a majority were not 
found, and there are various possible ways of 
determining a top-two (or sometimes top three) from a Bucklin election.)